Tumor cells express receptors for growth factors and cytokines that stimulate proliferation of the cells. Antibodies to such receptors can be effective in blocking the stimulation of cell proliferation mediated by growth factors and cytokines and can thereby inhibit tumor cell proliferation and tumor growth. Commercially available therapeutic antibodies that target receptors on cancer cells include, for example, trastuzumab which targets the HER2 receptor (also known as ErbB2) for the treatment of breast cancer, and cetuximab which targets the epidermal growth factor receptor (EGFR, also known as HER1 or ErbB1) for the treatment of colorectal cancer and head and neck cancer.
Monoclonal antibodies have significantly advanced our ability to treat cancers, yet clinical studies have shown that many patients do not adequately respond to monospecific therapy. This is in part due to the multigenic nature of cancers, where cancer cells rely on multiple and often redundant pathways for proliferation. Bi- or multi-specific antibodies capable of blocking multiple growth and survival pathways at once have a potential to better meet the challenge of blocking cancer growth, and indeed many of them are advancing in clinical development. However, bispecific antibodies present significant design challenges, due to the greatly increased number of variables that need to be considered in their design and optimization, as well as to their structural differences from naturally occurring antibodies.
Monoclonal antibodies such as trastuzumab, cetuximab, bevacizumab and panitumumab have significantly improved patient outcomes in the clinic, and over two hundred therapeutic monoclonal antibodies are currently being tested in clinical development. However, it has become apparent that tumors driven by single oncogenes are not the norm, and treatment often results in activation of resistance mechanisms, which in turn also require targeted intervention. For example, in multiple pre-clinical models of trastuzumab resistance, inhibition of IGF-1R restores sensitivity to trastuzumab. Combination of targeted agents has been attempted in the clinic, but so far they have had limited clinical success and in combination can be prohibitively expensive. The need to inhibit multiple targets either due to resistance or to tumors being driven by multiple growth factor pathways has led to increased interest in bispecific antibodies. Currently developed bispecific antibodies were typically designed in empirical fashion. Further, the pharmaceutical properties of these bispecific antibodies were almost invariably inferior to those of monoclonal antibodies. These factors present a major challenge to the development of bispecific anti-cancer therapies. Significant added benefit from targeting multiple cancer survival pathways can be derived from increased work up front to identify and engineer a bispecific antibody with optimal characteristics. This requires an iterative approach consisting of computational simulation to identify optimal targeting strategies and design specifications, engineering of inhibitors that possess these characteristics, and experimental validation of the therapeutic hypothesis. We separate this engineering framework into two categories: selection of appropriate molecular format with robust pharmaceutical properties and computational simulation to identify the best targets and optimal therapeutic design characteristics, e.g., in an IgG-like bispecific antibody (FIG. 8).
One of the main advantages of antibodies is their ability to bind tightly to virtually any extracellular target. This property is driven by two features of antibody variable regions (VRs): the large flat surface of six complementarity determining regions (CDRs) and the fact that antibodies have two binding arms that can target two molecules simultaneously. In bispecific antibodies, dual target binding results in tighter affinity because once one arm of the antibody is bound to an extracellular target, the second arm is restricted to a narrow region above the plasma membrane (about 100 angstroms), and is, therefore, concentrated near the cell surface. This results in a much faster secondary binding event that is not limited by diffusion. The acceleration of a secondary binding event is called. Both affinity and avidity are rationally engineerable properties, as the former can be improved via in silico affinity maturation and the latter can be enhanced by engineering additional targeting arms to the same or different antigen present on the cell surface. In addition to their binding capabilities, antibodies may possess multiple effector functions mediated by their Fc domain: antibody-dependent cellular cytotoxicity (ADCC) and antibody-dependent cellular phagocytosis (ADCP) that in humans are determined by interactions with activating FcγRI, FcγRIIa/c, FcγRIIIa and inhibitory FcγRIIb receptors; complement-dependent cytotoxicity (CDC) that is triggered by antibody binding to the components of the complement system; and long half-life that is mediated via active recycling by the neonatal Fc receptor (FcRn). All of these functions can be tuned to optimize the effectiveness of an anti-cancer therapy and are may be retained to advantage in a bispecific protein.
The variable fragment (Fv), composed of the variable heavy (VH) and variable light (VL) domains of an IgG antibody, is a minimal antibody fragment that displays full antigen binding. These variable domains can be successfully fused into a single chain construct (scFv), although affinity is often reduced to some extent compared to a whole native antibody. A majority of current bispecific formats feature one or several scFv modules attached to the N- or C-terminus of an IgG heavy chain or IgG light chain via a low complexity linker. Another bispecific antibody format is dual variable domain immunoglobulin (DVD-Ig). DVD-Ig consists of a first IgG heavy chain with a second VH domain linked to its N-terminus by a short linker, and a first IgG light chain with a second VL domain similarly linked to its N-terminus. The second VH/VL domains form a pair with specificity for one antigen while the first VH/VL form a separate binding site with specificity for a different antigen.
Bivalent formats of IgG-like antibodies have one potential limitation; they can cross-link cell surface antigens, some of which are activated by dimerization, triggering undesirable signaling events in an uncontrolled manner. To address this challenge, tunable monovalent bispecific antibody formats have been developed. MetMab, a one armed anti-c-Met therapeutic antibody created by incorporating asymmetric “knobs and holes” into the Fc fragment, has been shown to be effective in models of pancreatic cancer and is being investigated in multiple clinical trials. This “knobs and holes” format has been recently extended to incorporate an antibody fragment targeting EGFR, giving rise to a functionally monovalent bispecific protein targeting EGFR/ErbB1 and c-Met/HGFR. Gunasekaran et al. have described an alternative implementation of the “knobs and holes” concept by engineering complementary charged surfaces into the Fc fragment. Davis et al. have described a “SEED” approach that used modified asymmetric Fc containing fragments from human IgG and IgA to form heteromeric monovalent antibodies. Finally, Bostrom et al. has described a novel engineering approach to construct bifunctional Fab fragment that can bind either HER2 or VEGF with high affinity. When combined in a canonical antibody molecule, these Fab fragments will engage HER2 or VEGF with different valences that will depend on the cellular environment and growth factor concentration.
Another important component of a bispecific antibody design is the optimization of pharmaceutical properties. To be clinically useful, a therapeutic protein must be stable, remain soluble over an extended period of time and possess a robust manufacturability profile. Bispecific antibodies are typically less stable than monoclonal antibodies, and initially may not possess adequate pharmaceutical properties for development. They can be stabilized through molecular engineering, through downstream formulation activities, or, as most commonly practiced, through the combination of both approaches.
The importance of minimizing chemical manufacturing and control liabilities in small molecule drug candidates has been long recognized and the rules to predict drug-likeness have been proposed. Many groups have used conceptually similar approaches to assess the fitness of IgG based proteins by evaluating unfavorable sequence features such as: non-canonical disulfides or unpaired cysteines, extra glycosylation sites, tyrosine sulfation motifs, solvent accessible methionines, asparagine deamidation motifs, and acid cleavage sites. Extra glycosylation sites and asparagine deamidation sites are quite common features of natural antibodies sequences. In fact over 20% of variable domains of heavy chains are reported to be glycosylated and over 5% of germline genes contain asparagine-glycine deamidation motifs. Deamidation rates in antibodies can be reliably estimated using the method proposed by Robinson that suggests that structurally constrained loops do not form a succinimide intermediate efficiently and therefore are stable.
While the canonical N-linked glycosylation motif (NXS and NXT, where X is any aa but proline) can be easily detected in antibody sequence, O-linked sites, which are liabilities in that they can also negatively impact pharmaceutical properties, are more difficult to recognize. Recently several O-linked modifications of variable domains of antibody light chains have been reported, mostly in the proximity of GS rich sequence motifs. Many approaches to improve the affinity and stability of a candidate protein at the discovery stage exist including structure-guided design, focused library screening and yeast display; therefore, we find it beneficial to remove such potential liabilities at risk in the early proof-of-concept proteins.
Other liabilities, such as aggregation and immunogenicity are more challenging from an engineering perspective. Not only are both multifaceted properties, but it is also very difficult to adequately evaluate them in small scale biochemical and biophysical testing and, therefore, they tend to first be detected late in development. Arguably the best approach to reduce antibody immunogenicity is through humanization. This approach has been extensively validated by a number of humanized antibodies that have been well-tolerated in the clinic. A recently proposed “superhumanization” approach introduces human germline sequences into the CDRs to yield ‘fully human’ antibodies. This approach requires each amino acid (“aa”) in the CDRs to be mutated in order to determine its contribution to antigen binding; the resulting antibody may be less immunogenic. Aside from sequence-based features, immunogenicity of antibodies and antibody-like proteins can be dependent on their aggregation stability. Interestingly, the “humanness” and the stability in an antibody module, such as scFv, can be co-engineered via a knowledge-based approach.
Engineering of protein antibody solubility is another daunting task, as the property is also a composite of several physico-chemical parameters. Nevertheless, a number of methods to combat insolubility have been proposed. Pepinsky et al. have used glycoengineering, isotype switching, and structure-guided mutagenesis, to increase the solubility of a monoclonal antibody. Chemamsetty et al. have described an unbiased approach to improving stability with respect to aggregation that relies on molecular dynamics simulations to calculate a parameter called surface aggregation propensity and applied this technique to introduce stabilizing amino acids (“aas”) in the aggregation-prone regions of antibodies. Interestingly, in their analysis of antibodies, aggregation prone regions often co-locate with functionally important regions that confer Fc receptor or antigen binding and, therefore, cannot be easily removed.
Such a coupling between molecular function and pharmaceutical properties in antibodies is common. It can significantly complicate optimization of the bispecific antibodies, as even larger parts of their sequences are located in the functionally important regions. Therefore, to enable successful engineering it is important to identify critical molecular functions and optimal design characteristics.
Computational Simulation to Identify Best Targets and Optimal Therapeutic Design Characteristics
Computational simulation is a valuable tool for guiding drug development decisions, both in the laboratory and in the clinic. Population pharmacokinetic modeling is a mature example of the use of models to optimize dose scheduling and clinical trial designs. For therapies with known targets it is possible to utilize computational simulation much earlier in the design process. Advances in multiplex, high-throughput quantitative protein measurement technologies have enabled observation of the complex dynamics that occur in cellular signaling networks. These data permit the creation of network models which capture the mechanistic behaviors of biological systems that are relevant to diseases such as cancer. By simulating potential therapeutics through network modeling it is possible to design more effective therapeutics at an accelerated pace and more accurately predict design parameters.
Traditionally, selection of a pharmaceutical agent for a targeted therapy begins with a known target which is selected from a multitude of molecular, biological and physiological data. However, even in well-studied and heavily targeted biological systems there are opportunities for new discoveries, which can be aided by simulation of pathway or network models.
Accordingly, additional therapeutic approaches for cancer treatment, and in particular polyspecific antibody-based proteins that are engineered to have superior biophysical and therapeutic properties are difficult to obtain, but are needed to overcome limitations of current antibody therapies and to provide other benefits.